Industrial 3D printing, additive manufacturing (AM), has hit the bottom of the Gartner Hype Cycle. Many people are giving up hope that it will ever come back. Hasn’t it been overblown, and is additive manufacturing maybe just a tool for niche applications?
Not at all.
Industrial 3D printing has enormous potential, and it will profoundly change our physical world. But it needs to grow up. It has to get away from boutique manufacturing and prototyping. Many people, including me, have been saying this for years, and it remains true today.
But nobody is acting on this. I am shocked to see widespread confusion about the core of the issue.
There are really just two challenges holding AM back. Both are complex — both are solvable.
- Few people know how to design for additive manufacturing. The tools for engineers are immature and still seem exotic. People generally resist change unless inevitable. As a result, few designs exist that make sense for AM. And there is no urgency, because of point 2.
- Even if you design something for AM, the cost for production of the part is prohibitive. So, why bother?
I spent the better part of the last decade in an attempt to help with the first point, and I am continuing on that journey.
But let’s talk about the second point.
Why is AM expensive?
AM is one of the most expensive manufacturing processes today.
Is this justified?
To answer that, let us zoom out for a second: How does AM fundamentally compare to conventional manufacturing?
Traditional production lines require complex tooling, purpose-built machines, laborious manual assembly steps. They are inflexible and require substantial upfront investment. Changing and improving the manufactured products is hard, because the factory needs to undergo costly reconfiguration.
Additive manufacturing, on the other hand, is flexible and reasonably standardised. AM allows for the production of parts that are functionally integrated, requiring less human labour in assembly. While AM cannot produce everything, it should at least be able to replace several steps in a conventional factory.
Ask someone outside the industry which one is the more expensive process? The answer would be clear.
We all know that additive manufacturing for end production needs streamlining and improvements, more automation, and integration with other processes. But a resulting digital production line, with AM at is core, should be vastly cheaper than a conventional one.
So, why is AM more expensive?
The lopsided economics of 3D printing
Most of the objects we create at LEAP 71 have additive manufacturing at their heart. We constantly work with industrial 3D printing providers all over the world to check whether the resulting product makes financial sense. Often, at first glance, it doesn’t. But our main hypothesis says: if the product is superior and has a large market, there will be a way to produce it economically.
Get your FREE print subscription to TCT Magazine.
Exhibit at the UK's definitive and most influential 3D printing and additive manufacturing event, TCT 3Sixty.
But not now, not today — because of the strange economics of the AM industry.
As an example — to print a metal object in a 30cm cubed build volume, say an advanced heat exchanger, costs between $5,000 to $50,000, depending on the material. And the common wisdom is, that’s the way it is — sorry. 3D Printing is expensive. Serial production makes no sense — unless you want tiny parts in high-value application s.
I don’t subscribe to this argument.
If you look at any conventional product and its manufacturing process, you will be shocked how many steps are involved, the number of specialised machines which need to be built, and the amount of human-driven assembly required.
I recently watched a video that shows the manufacturing of a Christmas-tree-shaped cookie cutter. It involves the use of an enormous bespoke contraption of a machine to produce the object. Impressive engineering. But if you want to manufacture anything even slightly different, say, a star-shaped cookie cutter, someone needs to build a new device: A star-shaped-cookie-cutter-making system is substantially different from the Christmas-tree-shaped-cookie-cutter-making system. You can marvel at the ingenuity that went into their design — but can’t help but be amused that it should be necessary.
Visit any traditional factory and marvel at a specialised machine that bends a metal wire into a certain shape (it’s the only thing it does). A refrigerator-sized device which flips over a part for assembly in a later step. A robot exercising a surprisingly elaborate dance just to add a spot-weld. And sprinkled in between are humans hard at work with tasks like using a rubber hammer to flatten an electric motor coil, before it moves on to the next production step...
Conventional wisdom, starting with Henry Ford, says it’s fine to design such a bespoke production line, because you will be producing large amounts of these identical objects forever.
But it is a substantial undertaking, requiring a large upfront investment of both time and money. And forever is a long time, considering the pace of change in the world.
Contrast that with additive manufacturing: The same machine can produce a variety of products, and, because of its ability to create complex shapes, you can build functionally-integrated parts. You can skip many laborious assembly steps. By using AM, we should be able to reduce the complexity of a factory significantly.
Intuition would say: Using a standardised 3D printing system to build objects that are closer to the end product (because pre-assembled), should be cheaper than using specialised machinery that inflexibly can produce only one thing, and which requires additional human labour to assemble the final product. Intuitively, it should also be obvious that such a factory could be built faster and would require less capital investment.
A standardised additive manufacturing production cell can replace a large part of a bespoke assembly line, and should, therefore, be wildly economically successful.
Why is nobody building it?
Large AM providers are stuck in their small-volume, high-margin business way of thinking. They forget that the real market is the 99.9% of industry that is locked in the insanity of creating bespoke one-off-manufacturing lines for the gazillion products we buy every day.
Let’s talk about margins
One of my industry friends told me the story about the challenges they had, when ramping up the production of the first unibody aluminium notebook computers. Nobody could supply the massive number of CNC mills required to outfit the production lines. As a consequence, the first thing they had to do — before building a computer factory — was to create a factory that produced CNC mills at scale. The CNC mill production lines had to be automated to an unheard-of degree, because of the fast ramp up required. In a short period of time, starting from zero, they had to produce more CNC mills than the entire world’s existing output combined.
The first unibody notebook computer made absolutely no economic sense — until someone decided to break through the chicken-and-egg issue of the economics of CNC milling a consumer product at scale.
It would have been easy to say, it’s not economical, and the production capacity doesn’t exist. Let’s not do it.
But they did.
I keep this story in mind when I think about the production of 3D printers today. Automation is minimal, these machines are built by hand, over the course of weeks. As a result, each machine is rare and expensive. They are put into a shiny case and are sold like luxury cars, one by one.
But lately, profit margins have come under pressure, because of competition in the tiny and overcrowded market of AM.
This causes many of the recent troubles in the AM industry. More automation in the manufacturing of the printers could help, but requires upfront investment that is hard to come by in this challenging environment.
This is the root of the death spiral we are seeing in the AM market right now.
Using my friend’s example as guidance, here is an idea for industrial 3D printer companies: Eat your own dog food. Nobody will use your systems in serial production unless you do it yourself! Challenge your teams to build a fully automated manufacturing line for your AM machines with 3D printing at its core. Use the process to push into a different market, the market of integrated digital production. Build a machine that builds the machine. Use AM to save production steps. Use the project as a vehicle to make AM compatible with automation and integrate with conventional production steps.
What would happen if a printer manufacturer said, let’s build a standardised AM production cell for close-to-end-use parts, and, as the first application, produce the AM cell itself?
By thoroughly automating the manufacturing of 3D printers, AM companies could enter a virtuous circle which allows them to cut costs as a first step. But then, more importantly, vastly expand the market by lowering prices and being compatible with the requirements of an automated factory. You cannot do this incrementally, though. You have to make a decision, and then apply the willpower and money to make it happen.
Vision and execution.
Today, 3D printers are built like a Rolls Royce car — by hand. No wonder both products have similar price points.
But the cost of printers is not the biggest issue. The difference between the $5,000 and the $50,000 for the example metal part, which I cited at the beginning, is the cost of consumables.
Markups on 3D printing materials are outrageous when compared to the raw material cost. The argument is the same as with the printers: as long as the market is tiny, only small batches can be sold. It’s fine to send 5 kilos of powder via air freight from Canada and charge a fortune — how else would you do it? There are no economies of scale.
But even with that valid argument, markups are considerable. What would the markup have to be, if you sold 10 tons of aluminium powder, instead 10 kilograms? What about 100 tons? If you charged less, how much would your market expand? Again, it’s a chicken-and-egg problem.
Which leads me to the final step in the cascade: The companies providing printing services. I understand their math — but if you compare it to any other industry, their prices are fantastical.
The depreciation period for an AM machine is typically set to 5 years, which is significantly shorter than the 10-12 years or more commonly seen in other industries. That alone adds hundreds or even thousands of USD to every print job, because, in this way of accounting, a fifth of the capital investment needs to be recovered every year.
To pay for the extensive manual labour going into maintaining and supervising the machines, the time spent buying small batches of materials, swapping powders out, cleaning the printers, and all the other things humans do manually in a boutique service provider, service providers’ markups on consumables are 100s or even 1000s of percent.
A kilogram of metal powder, already expensive because of the pricing of the materials provider, now is marked up again until it reaches cost levels that are unrealistic for all but the highest-value applications.
A kilogram of copper powder ends up in the range of gold bullion. In a way, I guess, this makes it consistent with a Rolls-Royce-priced printer!
Because AM printing services are expensive, few customers use them. Consequently, the markups have to be even higher to cover the low utilisation of the printers, which in turn makes it more expensive — resulting in even less customers and killing entire fields of applications. A startup working on a breakthrough product requiring lots of iterations using AM? Prepare to raise millions just to pay for the printing.
If you compare the markups used in the AM industry to conventional manufacturing, it’s beyond ridiculous. We are in luxury brand territory.
But I am not saying people are making a fortune — quite the opposite. Many service providers are struggling to get by in this tiny market. But the current sorry state of additive manufacturing is a self-inflicted wound. Looking at AM vs. traditional manufacturing from this perspective should be a wake-up call for everyone.
If you want to know what’s broken, that’s it.
If things are costly, few people can afford them, and almost all AM use cases end up making no sense. The demand for expensive manufacturing is small, resulting in a tiny market for 3D printed goods. Low volumes require higher markups, which fuel the high costs, which lie at the root of the problem. And the death spiral continues. A tiny market results in the cut-throat competition which has killed many a company in the field of AM in the past few years.
The way out is taking a bold view of the future, and investing in that vision. Many manufactured goods would benefit from AM components. Few can afford them today. The market is gigantic, even if the end-products were slightly more expensive as a result.
But looking at the complexity of conventional manufacturing, goods with AM components should actually be cheaper.
More standardised processes, less parts, less human labour for assembly. From this perspective, it is absurd that 3D printing is more expensive. The fundamental market for AM is easily a significant double digit percentage of today’s conventional manufacturing market — and not the tiny fraction it is today.
We will not get there unless someone untangles the knot.
Breaking the circle
How do we escape the curse of small market economics keeping the 3D printing market tiny and holding back technological innovation which would benefit us all?
From my side, and our company, we have decided to ignore the economics altogether in the first analysis — if at all possible. And more companies should follow suit. The push for innovative digitally produced products will hopefully be the catalyst we need to enter a virtuous cycle.
Many of the products we design for our customers at LEAP 71 have significant market potential. If the size of the end market is large enough, and the object cannot be produced in a conventional factory, then we are certain, together with our customers, we will find a way to mass-produce it with industrial 3D printers at reasonable cost.
The market for CNC machines expanded enormously when the manufacturing of a modern computer required milling it from a block of steel. Someone did the math, made the investment, and it happened.
If the only way to produce a modern electric motor or air conditioning system is to build a fully integrated AM factory, someone will do it — and will put a metal atomizer next to it that produces powder at moderate prices.
But wouldn’t it be great if more people realised the potential and moved forward, driving the vision in parallel? Many applications would spring up that we cannot imagine right now.
One country which has realised that is China.
3D printing the same metal object we talked about earlier with Chinese providers costs between 50% and 90% less than using a AM provider from the US or Europe. Chinese industrial 3D printers are already competitively priced when sold in the West. They are even less expensive inside their home country — much closer to the markups and costs that I would consider healthy, using all the arguments I outlined above. The same is true for consumables.
I predict, as a consequence, driven by these more favourable economics, that we will see 3D printing for end-user parts being adopted in China much faster than anywhere else. This is good for humanity because we need these innovative products which can only be built with AM.
But it’s something for other 3D printing companies to consider — and for anyone who is involved in the industrial strategy of a country.
In conclusion
The economics of additive manufacturing are that of a tiny crowded market. The best way to break out would be for industrial 3D printer companies to build automated digital manufacturing lines for their own systems. This would also drive the integration into more industrialised workflows and bring down cost, which would create a virtuous cycle of adoption.
Even with existing AM technology, and expensive consumables, there is likely a market for service providers, who produce with much lower markups but higher machine utilisation, enabling AM applications that are currently too costly to consider. A service provider, producing only parts from one type of metal (to avoid downtime), with utilisation-driven prices would very likely be economically viable and a strong enabler for 3D printed end-use parts.
Everyone decrying why so many things around us are Made in China should consider why. Some of the most advanced manufacturing lines are located there because of strategic long-term investments made many years ago. Tim Cook famously remarked that the iPhone could not be built anywhere else in the world.
We are now seeing the need for the same type of investment in the field of additive manufacturing, and maybe we can copy China, for our benefit.
About the author
Lin Kayser is a serial entrepreneur and co-founder of LEAP 71, a company focused on developing computational and AI models for engineering.
Kayser's career spans three decades of building ventures in industrial automation, image processing and, with his previous company Hyperganic, design for additive manufacturing.
He and his partner, LEAP 71 co-founder Josefine Lissner, live in Dubai.
